A critical security flaw in Spring AI allows unauthenticated remote attackers to perform query injection, potentially leading to unauthorized data access or manipulation.

This vulnerability involves improper control of code generation (CWE-94) within FilterExpressionConverter implementations. Attackers can provide malicious input that is not properly escaped, allowing them to alter the intended query structure sent to vector stores without requiring authentication.

The ability to manipulate vector store queries poses a significant risk to data integrity and confidentiality. An attacker could potentially extract sensitive data, bypass intended access filters, or disrupt application availability, justifying the CVSS score of 8.6. This impact is compounded by the fact that the vulnerability is exploitable remotely by unauthenticated users, increasing the likelihood of widespread abuse.

Immediate Action: Update Spring AI to version 1.0.6 or 1.1.5 immediately to incorporate the necessary input escaping fixes.

Proactive Monitoring: Review application logs for suspicious query patterns or unexpected filter expressions that deviate from standard application behavior.

Compensating Controls: Deploy a Web Application Firewall (WAF) with rules configured to detect and block common injection patterns in API requests, though this should be considered a temporary measure until patching is complete.

Public Exploit Available: exploit_available (unknown)

Given the severity of this query injection vulnerability and the ease of exploitation, organizations using Spring AI must prioritize upgrading to the patched versions. Failure to remediate could allow unauthorized entities to compromise the underlying vector store data. Apply the update as soon as possible to ensure the security and integrity of your AI-driven infrastructure.